Biology and ecology of bigheaded carp in an invaded ecosystem

Transcription

1 Purdue University Purdue e-pubs Open Access Dissertations Theses and Dissertations Spring 2015 Biology and ecology of bigheaded carp in an invaded ecosystem Alison Adele Coulter Purdue University Follow this and additional works at: Part of the Natural Resources and Conservation Commons, and the Terrestrial and Aquatic Ecology Commons Recommended Citation Coulter, Alison Adele, "Biology and ecology of bigheaded carp in an invaded ecosystem" (2015). Open Access Dissertations This document has been made available through Purdue e-pubs, a service of the Purdue University Libraries. Please contact epubs@purdue.edu for additional information.

2 Graduate School Form 30 Updated 1/15/2015 PURDUE UNIVERSITY GRADUATE SCHOOL Thesis/Dissertation Acceptance This is to certify that the thesis/dissertation prepared By Alison Adele Coulter Entitled BIOLOGY AND ECOLOGY OF BIGHEADED CARP IN AN INVADED ECOSYSTEM For the degree of Doctor of Philosophy Is approved by the final examining committee: Reuben R. Goforth Chair Tomas Hook Jon Amberg Bryan C. Pijanowski To the best of my knowledge and as understood by the student in the Thesis/Dissertation Agreement, Publication Delay, and Certification Disclaimer (Graduate School Form 32), this thesis/dissertation adheres to the provisions of Purdue University s Policy of Integrity in Research and the use of copyright material. Approved by Major Professor(s): Reuben R. Goforth Approved by: Rado Gazo 4/14/2015 Head of the Departmental Graduate Program Date

3

4 i BIOLOGY AND ECOLOGY OF BIGHEADED CARP IN AN INVADED ECOSYSTEM A Dissertation Submitted to the Faculty of Purdue University by Alison Adele Coulter In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy May 2015 Purdue University West Lafayette, Indiana

5 ii ACKNOWLEDGEMENTS I would like to thank my advisor, Dr. Reuben R. Goforth for his help, encouragement and caring throughout the completion of this degree. He made my experience at Purdue University both exciting and productive. Additionally, I would like to thank Dr. Tomas Höök, Dr. Bryan Pijanoskwi, and Dr. Jon Amberg for agreeing to serve on my committee and for their recommendations and advice. Thank you also to Dr. Pijanowski, Doug Keller of the Indiana Department of Natural Resources, and Dr. Goforth for securing funding. Funding for this project was provided by the Indiana Department of Natural Resources, the Great Lakes Restoration Initiative, the Graduate Assistance in Areas of National Need Fellowship, the Ross Fellowship through the Department of Forestry and Natural Resources at Purdue University and the USGS Upper Midwest Environmental Sciences Center. I would also like to thank those who have contributed as co-authors on various projects including Doug Keller (Chapter 3 & 4), Elizabeth Bailey(Chapter 3 & 4), Heidi Swanson (Chapters 5 & 6), Jon Amberg (Chapter 6), and S. Grace McCalla (Chapter 6). Additionally, I would like to thank the Goforth and Höök Labs at Purdue University for helping to improve various manuscripts. Special thanks to David Coulter and Zach Feiner for providing additional help that improved the chapters in this dissertation.

6 iii This project involved extensive field work across four years and could not have been completed without the assistance of employees, volunteers and collaborators. Field and lab assistance was provided by Elizabeth Bailey, Allison Lenaerts, Courtney Cripe, Austin Prechtel, Jessica Leet, Wesley Goldsmith, Sam Nutile, Kensey Thurner, Preston Stipe, Tess Thoren, Megan Gunn, Jake Adams, Sara Andree, Jayson Beugly, Jing Tan, Conor Keitzer, Caleb Rennaker, Connor Emenhiser, Anthony Escobedo, Katherine Touzinsky, and other volunteers and employees of the Goforth Lab. Additional thanks go to numerous volunteers from the Purdue University student chapter of the American Fisheries Society. Field assistance was also provided by Tom Stefanavage, Nate Nye, and Craig Jansen of the Indiana Department of Natural Resources. Additional data were supplied by Sarah Huck and the Colombo Lab at Eastern Illinois University. Special thanks go to Steve of the Americus boat ramp for allowing us to use his private boat ramp at no cost. Finally, I would like to thank my family, friends and lab mates who were all invaluable to the completion of this research. Members of the Höök and Goforth Labs provided important feedback on manuscripts and presentations. Thanks to Tim Sesterhenn for advice. Thanks to my parents, Jeanne and Gary Strandburg, for their continued support during this Ph.D. and throughout my life. Special thanks go to my husband, David Coulter. Thank you to everyone.

7 iv TABLE OF CONTENTS Page LIST OF TABLES... vii LIST OF FIGURES... x ABSTRACT... xiii CHAPTER 1. INTRODUCTION... 1 CHAPTER 2. SILVER CARP POPULATION CHARACTERISTICS IN THE WABASH RIVER (INDIANA, USA): VARIATION IN SIZE STRUCTURE, RELATIVE WEIGHT, SEX RATIO, AND LENGTH-WEIGHT RELATIONSHIPS Abstract Introduction Methods Temporal variation Spatial variation Sex ratio Length-at-maturity Results Temporal variation Spatial variation Sex ratio Length-at-maturity Discussion CHAPTER 3. INVASIVE SILVER CARP MOVEMENTS IN THE WABASH RIVER (INDIANA, USA), A PREDOMINANTLY FREE-FLOWING MIDWESTERN RIVER... 27

8 v Page 3.1 Abstract Introduction Methods Surgery Genetics Telemetry Movement Trends Model Selection Results Movement Trends Model Selection Discussion CHAPTER 4. PREDICTORS OF BIGHEADED CARP DRIFTING EGG DENSITY AND SPAWNING ACTIVITY IN AN INVADED, FREE-FLOWING RIVER Abstract Introduction Materials and Methods Results Discussion CHAPTER 5. HYBRIDIZATION BETWEEN TWO INVASIVE FISHES AND ITS IMPLICATIONS FOR THEIR MANAGEMENT Abstract Introduction Materials and Methods Collection and tagging of adults Genetic identification of hybrid adults and eggs Length-weight regressions, lengths and condition Determination and analysis of movements Egg size and abundance... 88

9 vi Page Stable isotope analysis Results Movements of bigheaded carp Egg size and abundance Stable isotope analysis Discussion Conclusions CHAPTER 6. SEASONAL RESOURCE OVERLAP OF INVASIVE BIGHEADED CARP AND NATIVE FISHES USING STABLE ISOTOPES Abstract Introduction Methods Sample collection Stable isotope analysis Results Discussion LITERATURE CITIED VITA

10 vii LIST OF TABLES Table... Page Table 2.1 Catch, total length (L T ±SE), and relative weights (W R ±SE) of Silver Carp collected from from the Wabash River. All year employed boat electrofishing (EF) while Silver Carp were also collected in 2013 using gill netting (GN), hook-and-line sampling (HL), and fyke netting (FN). Data were also collected in 2013 from Silver Carp that jumped into boats (JB). The river kilometer (rkm) of collection if also indicated Table 2.2 Numbers of male and female Silver Carp collected via electrofishing from river kilometer 499 of the Wabash River. Only months where sex was determined for > 3 individuals are shown. No Silver Carp were sexed during Table 2.3 Length-weight regressions for male and female Silver Carp using weight (W, g) and total length (L T, mm). Individuals collected in 2010 were not sexed. All regressions were significant (P < ) Table 3.1 Summary of detections collected using Vemco VR2W stationary receivers and active tracking of Silver Carp including the mean number of detections fish -1 year - 1 for Silver Carp. The number of VR2Ws are the number of stationary receivers deployed during that year Table 3.2 Results of model selection procedures to predict movement distance of Silver Carp using a linear mixed model. Only models with delta AIC < 2.00 are shown. Variables included in the model were: river gage (Gage), change in gage over 24 hour (Gage24), growing degree day (GDD), cumulative growing degree day (CGDD), day of year (DOY), year, month, and sex. Year and sex were categorical. Individual fish were included in all models as a random effect. Information presented on each model includes AIC with small sample size correction (AICC), delta AIC ( AIC), AIC weight (WAIC) and cumulative AIC (CAIC)... 54

11 viii Table... Page Table 3.3 Results of model selection procedures to predict movement rate of Silver Carp using a linear mixed effects model. Only models with delta AIC < 2.00 are shown. Variables included in the model were: river gage (Gage), change in gage over 24 hour (Gage24), growing degree day (GDD), cumulative growing degree day (CGDD), day of year (DOY), year, month, and sex. Year and sex were categorical. Individual fish were included in all models as a random effect. Information presented on each model includes AIC with small sample size correction (AIC C ), delta AIC ( AIC), AIC weight (WAIC) and cumulative AIC (CAIC) Table 3.4 Results of model averaging of all models with ΔAIC < 2. Average model coefficients are listed with associated p-values listed below in italics. Models for movement distance (m) and movement rate (m/week) were linear mixed effects models while the model for movement probability (move = 1, stationary = 0) was a logistic mixed effects model. Variables included in the model were: river gage (Gage), change in gage over 24 hour (Gage24), growing degree day (GDD), cumulative growing degree day (CGDD), day of year (DOY), year, month, and sex. Year and sex were categorical. Individual fish were included as random effects Table 3.5 Results of model selection procedures to predict movement probability using a logistic mixed effects model. Only models with delta AIC < 2.00 are shown. Variables included in the model were: river gage (Gage), change in gage over 24 hour (Gage24), growing degree day (GDD), cumulative growing degree day (CGDD), day of year (DOY), year, month, and sex. Year and sex were categorical. Individual fish were included in all models as a random effect. Information presented on each model includes AIC with small sample size correction (AIC C ), delta AIC ( AIC), AIC weight (WAIC) and cumulative AIC (CAIC) Table 4.1 Summary of bigheaded carp spawning activity in the Wabash River, IN, USA. Spawning descriptors include dates of initial egg detection (initial), peak egg density (peak) and the last day eggs were detected that year (end). The maximum and minimum water temperatures at which eggs were detected, the maximum and minimum egg densities collected each year, and the cumulative growing degree day (CGDD, base 10 C) at initial egg detection (initial) and at peak egg density (peak) are provided Table 4.2 Best models as indicated by backward model selection using penalized likelihood ratios for the initiation and cessation of spawning. Coefficients in the best models are listed with associated confidence intervals (C.I.) and p-values. Models included day of year (DOY) and cumulative growing degree day (CGDD, base 10 C). Other possible coefficients included in full models but in none of the final models were amount of daylight (hrs), water temperature ( C), river gage (m), and change in river gage height over 24 hours... 77

12 ix Table... Page Table 4.3 Results of predictive models for 2014 bigheaded carp spawning season in the Wabash River, IN, USA. Actual egg densities represent the mean of three bongo net pulls (±SE) except for day of year (DOY) 162 where only one bongo net pull was done. Models were constructed from 2012 and 2013 data. Initiation and cessation spawning models used penalized logistic regression to predict the probability that spawned eggs would be present at a given DOY, spawned eggs while the density model was a linear model to predict egg density (eggs/m 3 ). Model coefficients are listed in Table 4.2 and Table 4.4 Generalized linear models to predict drifting egg density (eggs/m 3 ) in the Wabash River, IN, USA, selected based on Akaike s Information Criterion corrected for small sample size (AIC C ). Full model included cumulative growing degree day (CGDD, base 10 C), day of year (DOY), river gage height, water temperature, daylight, and change in river gage over 24 hr Table 4.5 Results of model averaging to predict drifting egg density (eggs/m 3 ) in the Wabash River, IN, USA, of generalized linear models from Table 4. This averaged model exhibited an R 2 = Table 5.1 Summary of characteristics of bigheaded carp groups (i.e., bighead, silver, early hybrid and later hybrid) from the Wabash River, IN, USA. For lengthweight regressions, total length (TL) was in mm and weight (W) was in g which were log transformed for linearity Table 5.2 Mean (±se) stable isotope values of bigheaded carp groups (i.e., bighead, silver and later hybrid) from the Wabash River, IN, USA, collected across three seasons Table 6.1 Summary of δ 13 C and δ 15 N ( o / oo ) values and metrics for native (bigmouth buffalo, gizzard shad) and invasive silver-hybrid carp (silver and hybrid carp combined) and bighead carp from the Wabash River, Indiana, USA. Metrics include nearest neighbor distance (NND), total area of the convex hull (TA) and size of the Bayesian ellipses (SEA C ) Table 6.2 Seasonal percent overlap based on area of small sample size corrected Bayesian ellipses (SEA C ) overlap of silver-hybrid carp (silver and hybrid carp combined) with bigmouth buffalo and gizzard shad. Percentage was calculated relative to the area of the species listed as column headings (i.e., 0.9% of the bigmouth buffalo SEA C overlapped with silver carp SEA C

13 x LIST OF FIGURES Figure... Page Figure 2.1 Temporal variation is electrofished Silver Carp total length (L T ±SE) from river kilometer 499 in the Wabash River, Indiana, USA. Letters indicate significant differences Figure 2.2 Distributions of total lengths of electrofished Silver Carp from river kilmometer 499 collected from in the Wabash River, Indiana, USA Figure 2.3 Temporal variation is electrofished Silver Carp relative weight (W R ±SE) from river kilometer 499 in the Wabash River, Indiana, USA. Letters indicate significant differences Figure 2.4 Length-weight relationships for male and female Silver Carp from the Wabash River. Regression equations are listed in Table Figure 2.5 Mean total length (±SE) of electrofished Silver Carp captured at four locations in April and May 2011 from the Wabash River, Indiana, USA. Significant differences are indicated by the letter groupings Figure 2.6 Logistic regression of maturity (0 = immature, 1 = mature) of Silver Carp collected from river kilometer 499 of the Wabash River, Indiana, USA, from using all gear types Figure 3.1 Locations of Vemco VR2W stationary receivers within the Wabash River, Indiana, USA, indicated by triangles. Area included in the enlarged image is indicated by the black square. Area of river covered by active tracking is included in the shaded square. Scale bar is for enlarged area of Wabash River Figure 3.2 Hydrograph of the Wabash River, IN, USA from United State Geological Survey river gage #

14 xi Figure... Page Figure 3.3 Summary of the movements of Silver Carp by month. a) Proportion of consecutive detections indicating no movement. Of those individuals that did move, b) Mean (1/2SD) movement rate and distances varied by month. c) Directionality of Silver Carp movements upstream of downstream of their previous detection. Movements from January, February, November and December were not included as small sample sizes prevented accurate comparisons (n = 1, 2, 8, and 1 respectively) Figure 3.4 Movement distances and rates from across months included in twofactor blocked ANOVA (month and year) with an individual Silver Carp random effect. a) Movement distance and rate (1/2 SD) was greater in 2013 than in 2011 and Mean monthly b) movement distance (1/2 SD) and c) movement rate (1/2 SD) of Silver Carp from The highest movement rates for most months were in 2013 (> 2011 in June and August; > 2012 in May-July). September had significantly higher movements rates ( ) as did October (2011) than summer months (May-August). In 2012, March also had higher movement rates than summer months (May-August) Figure 3.5 Summary of the movements of Silver Carp in a) June and b) September. Months with higher movement rates tend to have more broad-scale movements while summer months tended to have more, smaller movements Figure 3.6 Number of tagged Silver Carp located and hydrograph at different locations on the Wabash River, Indiana, USA, in Days of year (DOY) included were selected to cover pre-spawn and spawn time periods. Shaded area in Fig 2a indicates time period enlarged in other figure parts (b-d). Locations of the stationary receivers (VR2Ws) whose data are displayed are: a) Backwater 1 located near River Kilometer (RKm) 499; b) Backwater 1; c) RKm510; d) RKm521. Dashed line is gage height and solid line is number of tagged fish Figure 4.1 Map of location on the Wabash River, IN, USA, at RKM 499 where sampling of bigheaded carp eggs occurred. Additionally, the positions of the weather station where temperature data was obtained for growing degree-day calculations is shown as well as the location of the river gage that measured discharge Figure 4.2 Egg densities collected across the A) 2012, B) 2013 and C) 2014 spawning seasons from the Wabash River, Indiana, USA. The solid line represents hydrograph while the dashed line indicates water temperature Figure 5.1 Abundance of bigheaded carp groups (i.e., bighead, silver and hybrid [early and later hybrids combined]) in adults and drifting eggs collected from the Wabash River, IN, USA... 99

15 xii Figure... Page Figure 5.2 Length-weight regressions for bigheaded carp groups (i.e., bighead, silver, early hybrid, later hybrid) collected from the Wabash River, IN, USA. Regression lines and statistics are reported in Table Figure 5.3 Summary of the movements (mean ±se) of bighead (n = 3), silver (n = 133), early hybrid (n = 5) and later hybrid (n = 10) bigheaded carps in the Wabash River, IN, USA. Movement distance and movement rates were average for each individual, and these values were used to calculate average values for each bigheaded carp group Figure 5.4 Abundances of silver, hybrid (early and later hybrids combined) and bighead carp drifting eggs collected across spawning seasons relative to the average number in each bigheaded carp group within each spawning season from the Wabash River, IN, USA. Dashed lines are at zero which indicates when observed drifting egg abundance of each bigheaded carp group is equal to the mean abundance within that spawning season Figure 5.5 Egg sizes (mean ±se) of bighead, silver and hybrid (early and later hybrids combined) carps from the Wabash River, Indiana, USA, in 2012 and Figure 5.6 Stable isotope values (mean ±se) for bighead (n=1), silver (n=22), and later hybrid (n=9) carp from the Wabash River, IN, USA. Means represent the combination of all three seasons (Table 5.2) Figure 6.1 Seasonal δ 13 C and δ 15 N mean values (se) for native species (bigmouth buffalo and gizzard shad) and invasive silver (silver and hybrid carp combined) and bighead carp from the Wabash River, Indiana, USA. Fall samples were collected in 2012 while spring and summer samples were collected in Periphyton and gastropods were collected as baselines in the fall and summer Figure 6.2 Silver-hybrid carp (silver and hybrid carp combined) mean seasonal δ 13 C and δ 15 N values (se) from the Wabash River, Indiana, USA

16 xiii ABSTRACT Coulter, Alison Adele Ph.D., Purdue University, May Biology and Ecology of Bigheaded Carp in an Invaded Ecosystem. Major Professor: Reuben Goforth. Globally, the homogenization of species has become a threat to biodiversity. As species are transported around the world, a portion of these species, released intentionally or accidentally, may become invasive and can produce negative impacts. Great effort has been invested into early identification and prevention of invasions as these are considered less expensive than managing an invasion. Unfortunately, species may exhibit varying characteristics across ecosystems, and so their behavior and potential survival in a new environment may be difficult to predict. Therefore, I examined trends in the biology and behavior of invasive fishes, including the plasticity surrounding these and how they may contribute to successful invasions, using bigheaded carps (Hypophthalmichthys spp., silver and bighead carp and their hybrids) in the Wabash River, Indiana (USA), as a case study. Trends in population characteristics appeared to vary with invasion stage. Female-skewed sex ratios, changing length-weight relationships, and earlier maturation are all characteristics that may ultimately contribute to the successful establishment of these fishes along invasion fronts. Movements could be extremely rapid but exhibited predictable patterns that may facilitate the management and control of these invasive fishes. Reproduction in these species was influenced by different environmental cues

17 xiv than those from their native ranges and was dependent on growing degree day rather than changes in river discharge. Hybrid bigheaded carp (silver x bighead carp) were not different from silver carp in any of the characteristics examined (i.e., movements, condition, diet) but are increasingly represented in adults and eggs in this system and may serve to increase heterozygosity. Stable isotope analysis indicated that there was little dietary overlap between bigheaded carps and native planktivores. Additionally, there was seasonal variation in resource use that may function to minimize this overlap. Overall, many of the variables examined in these studies may be influential in facilitating the successful establishment and spread of these invasive fishes.

18 1 CHAPTER 1. INTRODUCTION Global ecosystems are experiencing change and disturbance at unprecedented rates, and much of this disturbance is anthropogenic in origin. Progress in the rapid transport of goods worldwide has led to increased opportunities for exotic species to arrive safely in a variety of ecosystems after a relatively short trip (Perrings et al. 2005; Meyerson and Mooney 2007; Hulme 2009), resulting in biotic homogenization (McKinney and Lockwood 1999; Olden et al. 2004; McKinney 2006). Humans have also manipulated landscapes, both establishing connections that may allow species to move more freely (e.g., bridges, canals) and installing barriers to dispersal (e.g., fences, dams) which can influence species invasions (Didham et al. 2007; Galil et al. 2007). Many exotic species, either intentionally or unintentionally introduced, fail to establish in the receiving ecosystem into which they have been introduced. An even smaller proportion of exotic species become invasive, meaning that their population numbers become high, they outcompete native species to become a dominant presence in the ecosystem, and they have severe negative impacts (Williamson and Fritter 1996; Kolar and Lodge 2001; Lodge et al. 2006). To become invasive, non-native species progress through a step-by-step process, including transport, introduction establishment, spread, and impacts (Kolar and Lodge 2001; Lockwood et al. 2005). The most cost effective way to prevent and control the

19 2 impacts of invasive species is to prevent introductions from occurring (Leung et al. 2002; Lodge et al. 2006). However, the volume of species being transported globally can make this financially difficult, and so many countries utilize risk assessments and rating systems to determine which species are most likely to become invasive and are therefore of highest priority for prevention (e.g., Copp et al. 2005). Non-native species introduced into a novel environment may exhibit differences in life history traits or behavior compared to what is observed in the native range (e.g., Bossdorf et al. 2005; Beaumont et al. 2009; Deters et al. 2013). Observations of this variation can reveal phenotypic plasticity and adaptations which can positively impact the invasion success of a species (Davidson et al. 2011; Lande 2015). Examination of the biology and ecology of invasive species can provide insights into a variety of topics. Because of the potential for variation in the biological and ecological traits of invasive species and the potential utility of this information in risk assessments (Kolar and Lodge 2002; Leung et al. 2002; Anderson et al. 2004; Leung et al. 2012), it is important to continue to examine these traits in invasive species. Examination of invasive species case studies can help to predict the future ecological impacts of a broader range of invasive species. Additionally, information regarding how species progress through the invasion process can improve the understanding and management of native species and basic ecology (Lodge 1993; Sakai et al. 2001). Successful invasive species generally exhibit a suite of traits that contribute to their success in novel environments, including rapid maturation and growth (e.g., Fox et al. 2007; Amundsen et al. 2012) and high propagule pressure (e.g., Lockwood et al. 2005), among others. Invasive species are also typically tolerant of a wide range of

20 3 environmental conditions (e.g., Williamson 1996; Briand et al. 2004), including conditions unlike those they may experience within native ranges. Exposure to unique or novel environmental conditions in invaded ecosystems may produce unexpected responses from invasive species, resulting in highly variable traits among ecosystems. As a result of exposure to these novel environmental conditions, invasive species may also display plasticity in a variety of traits that ultimately allow their successful establishment under different conditions (Davidson et al. 2011; Lande 2015). Examination of variation in traits across spatiotemporal contexts can also increase understanding of the plasticity and limitations of species. Bigheaded carp (Hypophthalmichthys spp.) are fishes that are rapidly becoming global species of concern as they have successfully invaded across countries and continents (Kolar et al. 2007). Valuable food and aquaculture species in their native range, these fishes have been introduced both intentionally and unintentionally, including in North America s Mississippi River Basin (Kolar et al. 2007). The recent spread of these species through North America has allowed for examinations of their biology and ecology across highly varied river ecosystems. Most of the current research on bigheaded carp has occurred in highly regulated river ecosystems of Midwestern North America (but see Stuck et al. 2015). However, as these species continue to expand their ranges, they are likely to occur in smaller rivers with fewer locks and dams, and potentially greater environmental variability as a result of reduced regulation of flow and climate change. Therefore, the overarching goal of this work was to examine the biology and ecology of bigheaded carp in a relatively free-flowing, invaded river ecosystem and how the observed traits may vary temporally.

21 4 The first chapter examines how general population characteristics of invasive bigheaded carp changed temporally. Changes in energy allocation through invasion stages may allow species to more easily establish by promoting early and rapid reproduction (Copp et al. 2004; Fox et al. 2007). The next two chapters link environmental conditions with the movement and reproduction of bigheaded carp, including examinations of how variable movement and spawning may be across individuals or through time. These chapters also highlight the resilience of these species, as there was observed reproduction without environmental cues previously thought to be necessary. Despite the observed variation, there were some general trends observed related to movement and reproduction that may enhance predictions and modeling efforts. The fourth chapter examines the hybridization of two invasive species, which may enhance the invasion capabilities of the parental species (Petit et al. 2003; Fitzpatrick and Shaffer 2007). It also considers how hybrids may function differently from the parent species. The establishment of hybrids as a separate population does not appear to be occurring in the Wabash River, with hybrids largely functioning similarly to one of the parental species, silver carp. The final chapter examines how invasive bigheaded carp may impact competition for primary productivity resources through resource overlap. As a complete work, this dissertation showcases the plasticity and resilience of bigheaded carps to highly varied environmental conditions and provides insights into what traits may contribute to the successful establishment of these fishes and other invasive species.

22 5 CHAPTER 2. SILVER CARP POPULATION CHARACTERISTICS IN THE WABASH RIVER (INDANA, USA): VARIATION IN SIZE STRUCTURE, RELATIVE WEIGHT, SEX RATIO, AND LENGTH-WEIGHT RELATIONSHIPS 2.1 Abstract Invasive species are an issue of concern worldwide and have been linked to many deleterious impacts, including loss of global biodiversity. Silver Carp (Hypophthalmichthys molitrix) have successfully invaded ecosystems in numerous countries and have been linked to declines in native fishes. Population-level characteristics (e.g., size structure, length-weight relationships) may vary across ecosystems and through time. Assessment of these characteristics can aid in the understanding of how Silver Carp move through the invasion process and provide insights into how populations of these fish may vary annually. This study sought to examine population-level characteristics of Silver Carp in an invaded ecosystem, the Wabash River (Indiana, USA). Silver Carp were collected via electrofishing at one location in the Wabash River from Silver Carp were also collected at three additional sites during 2011 to allow for spatial assessments of population characteristics. Mean total length of Silver Carp decreased through time and in a downstream direction, although relative weight generally increased over time. Length-weight relationships for Silver Carp were different between males and females and across years. Female fishes generally exhibited steeper slopes in length-weight relationships than males, but this was

23 6 not consistent for all years. Observed sex ratio was skewed towards females or males depending on season or year. The smallest mature female and male Silver Carp were 410 mm and 413 mm total length, respectively, and 50% maturity occurred at 440 mm total length. Overall, the results of this study reveal the dynamic nature of invading populations of Silver Carp. Additionally, population characteristics of Silver Carp are understudied in free-flowing invaded rivers, and population characteristics of these invasive fishes in the Wabash River should inform monitoring and control efforts in other largely unregulated rivers threatened by Asian carp invasion. 2.2 Introduction While only a small percentage of non-native species become invasive (Williamson and Fitter 1996; Kolar and Lodge 2001), the large volume of introductions (intentional and unintentional) increases the likelihood that species invasions may occur. The high ecological and economic risks posed by invasive species have led to evaluations of the traits that contribute to a non-native species becoming invasive (e.g., Lee 2002; Marchetti et al. 2004; Alcaraz et al. 2005; Pyšek et al. 2012). However, many invasive species exhibit considerable variation in morphological, behavioral, trophic, and life history characteristics in invaded ecosystems (Corkum et al. 1998; MacInnis and Corkum 2000b; Sakai et al. 2001; Alcaraz et al. 2005; Fox et al. 2007; Russell et al. 2012; Coulter et al. 2013), and population age (Lorenzen and Enberg 2002) and invasion stage (Feiner et al. 2012) can produce additional variation in these traits. Moreover, population-level characteristics of invasive species can change through the invasion process (i.e., introduction, establishment, spread; Sakai et al. 2001; Kolar and Lodge 2002) and across ecosystems (Rypel 2014). Such variation in population characteristics in response to the

24 7 novel environmental properties of invaded ecosystems can make efforts to model and control invasive species spread especially difficult. Relatively few studies exist detailing how population characteristics of invasive species change temporally, spatially, and among populations (Bøhn et al. 2004). However, basic demographic and life history information is essential for managing invasive species populations (Sakai et al. 2001), especially given increased recognition of plastic responses of these species to novel environments of invaded ecosystems. Silver Carp (Hypophthalmichthys molitrix) are global aquatic invaders, occurring in >70 countries and territories (Kolar et al. 2007). This species exhibits many of the traits associated with highly successful invasive species, including high fecundity, large body size, and rapid growth. Models and risk assessments have sought to predict and evaluate where Silver Carp may become invasive (e.g., Andersen et al. 2004; Jiménez-Valverde et al. 2011); however, these efforts have been largely based on life history and ecological information from the species native range and for proxy species. In addition, Silver Carp are known to exhibit plasticity in life history traits similar to other invasive species (e.g., Coulter et al. 2013), although variation in other population traits is unknown. There is thus considerable need for studies of Silver Carp population characteristics in introduced ecosystems to better understand how these properties change in space and time. In the last decade, understanding of Silver Carp population characteristics in invaded ecosystems has improved for some traits. For example, length-weight relationships have been determined in North America for the Missouri River (Wanner and Klumb 2005), several of its tributaries (Hayer et al. 2014), the Mississippi River (Williamson and Garvey 2005), and the Illinois River (Irons et al. 2011). Size structures

25 8 have also been assessed (Phelps and Willis 2013; Hayer et al. 2014; Stuck et al. 2015), as have age and growth (Williamson and Garvey 2005; Hayer et al. 2014; Stuck et al. 2015). However, many of these studies have relied on data from highly regulated river systems and are often based on established, high-density populations (but see Hayer et al. 2014; Stuck et al. 2015). Therefore, there is need for additional basic information on moderate or low-density Silver Carp population characteristics in free-flowing rivers such as the Wabash River (Indiana, USA). Silver Carp have increased drastically in abundance since their appearance in both the Mississippi (Wanner and Klumb 2009; Kelly et al. 2011) and Illinois Rivers (Chick and Pegg 2001). However, densities of Silver Carp are not as high in the Wabash River (Edgell and Long 2009; Stuck et al. 2015) even though first detections of Silver Carp in the Illinois and Wabash Rivers were relatively similar (Illinois: 1998, Wabash: 2003; USGS 2012). One of the potential causes of lower densities of Silver Carp in the Wabash River may be higher flows and water velocities (Stuck et al. 2015). The Wabash River has one mainstem dam and over 600 km of free-flowing river. Silver Carp are known to prefer low velocity and backwater habitats (Kolar et al. 2007; DeGrandchamp et al. 2008; Calkins et al. 2012), and flow patterns can also influence Silver Carp reproduction (Kocovsky et al. 2012). Differences in both population density and flow regime in the Wabash River may produce different population characteristics in Silver Carp compared to invaded rivers that are highly regulated. Demographic and life history studies of Wabash River Silver Carp should thus help to inform management of other largely unregulated rivers that Silver Carp may invade in the future (e.g., Great Lakes and upper Mississippi River tributaries).

26 9 This study sought to evaluate spatiotemporal trends in Silver Carp population characteristics in the Wabash River as a model invaded free-flowing river. Specifically, size structure (e.g., mean total length [L T ], length-frequency trends), condition (e.g., relative weight [W R ]), sex ratio, and length-at-maturity were assessed for Silver Carp in the Wabash River. We hypothesized that Silver Carp size structure, length-weight relationships, and condition would change annually to reflect growth, reproduction, and annual climatic and resource shifts. Because Silver Carp are highly mobile (DeGrandchamp et al. 2008), we hypothesized that spatial variation in size structure and condition within a year would not occur. Additionally, we hypothesized that Silver Carp would exhibit a 1:1 sex ratio, and that length-at-maturity would be similar to other invaded ecosystems of similar climate (e.g., 600 mm; Costa-Pierce 1992). 2.3 Methods Temporal variation Silver Carp were captured from within 2 km of river kilometer (rkm) 499 in 2010, 2011, 2012, and 2013 between March and October using boat electrofishing. The boat electrofisher (Model SR16H; Smith-Root Inc., Vancouver, WA, USA) was powered by a generator with a pulsator running at either 3-4 A of direct current at 30 pulses/s and 20-50% range of pulse width or 7-8 A of direct current at 120 pulses/s. Some of the Silver Carp were euthanized with an overdose of MS-222 (Tricaine S) prior to handling. Other individuals were involved in a separate telemetry study and were anesthetized for handling using electroanesthesia. Weight (W, g) and total length (L T, mm) were measured, although equipment failure resulted in W not being available for all individuals in Sex was also determined visually. Relative weight (W R ) can aid in assessing the general

27 10 health of a population (Murphy et al. 1991). Therefore, W R was calculated to evaluate condition in Silver Carp using a standard weight (W S ) equation for Silver Carp generated from multiple populations across North America (J.T. Lamer, personal communication; Fulton s condition factor produced the same significant differences as W R ). Trends in L T, length-frequency, W R, and length-weight relationships were all used to evaluate temporal changes in the Silver Carp population. To statistically compare mean L T between years, L T of the electrofished Silver Carp from rkm 499 was used in an analysis of variance (ANOVA) with a post-hoc Tukey s test. Length-frequency distributions (10 mm bins) were created to assess size structure. Population trends in Silver Carp W R were compared among years with the same methods as L T, to evaluate changes in condition through time. To determine the length-weight relationship for Silver Carp, L T and W data from individuals captured at rkm 499 were log 10 transformed for linearity. Differences in length-weight regression between sex and among years were tested using analysis of covariance (ANCOVA) to determine if length-weight regressions could be combined between sexes or among years Spatial Variation Silver Carp were captured from three additional locations in the upper Wabash River (rkm 521, 565, 600) in April and May of 2011 via electrofishing. Silver Carp collected at all four locations were evaluated using ANOVA to determine whether spatial differences existed in L T, and W R. Insufficient data were available from males and females, and so length-weight relationships were not evaluated.

28 Sex ratio Sex ratio was calculated for sampling dates where sex was determined for more than three individuals. Silver Carp used in the telemetry study were not used in the determination of sex ratios as the need to keep the individual viable occasionally prevented determination of sex Length-at-maturation Silver Carp captured using electrofishing and additional gear types were used to determine length-at-maturation for Silver Carp in the Wabash River. Additional Silver Carp were collected at rkm 499 using gill nets, hook-and-line sampling, and fyke nets in 2013 to increase sample size for the length-at-maturation analysis. Gill nets (12.7 cm mesh, 30.5 m in length) were set for 1.5 hr and fyke nets (1 x 2 m mouth, 5 m lead, 10 mm mesh) were set overnight. Logistic regression was used to determine L T at which 50% of Silver Carp were mature. 80% of Silver Carp were randomly selected and used to create the logistic equation. The remaining 20% were used to evaluate model accuracy. The resulting length-at-maturation value was compared to climatically similar ecosystems. All statistics were run in R (v ) with an α = Results Temporal variation A total of 356 Silver Carp were captured via electrofishing at rkm 499 with 70, 58, 77 and 151 individuals collected for 2010, 2011, 2012, and 2013, respectively. Total lengths of fishes captured ranged from mm (X = mm ±5.5SE; Table 2.1). Silver Carp L T near rkm 499 decreased through time (ANOVA: F 3,355 = 17.45, P < ; Figure 2.1), with 2010 and 2011 having significantly greater mean L T than 2013.

29 12 Additionally, 2013 mean L T was significantly smaller than Length-frequency distributions of Silver Carp show smaller individuals were increasingly represented in the electrofishing catch (Figure 2.2). W R averaged 1.04 ± 0.01SE (n = 311) overall, indicating collected fish were in good condition. Mean W R was different among years (ANOVA: F 3,310 = 8.70, P < ; Figure 2.3), with 2010 having a significantly smaller W R than W R was similar to 2010 but was significantly smaller than 2011 and and 2012 W R were not significantly different from each other or from Analyses of length-weight relationships between sexes and among years indicated that both of these factors were significant (F 1,262 = 6548, P < , P sex = , P year = 0.003) as were their interactions with length (P sex*lt < , P year* LT = 0.004). This indicated differences in the length-weight relationship slopes and intercepts. Therefore, equations for sexes and years were reported separately (Table 2.2; Figure 2.4). All lengthweight regressions were significant (P < ). Small sample size did not allow for separation of pre- and post-spawn females into separate regressions Spatial variation Mean L T of Silver Carp captured via electrofishing in April and May of 2011 at four sites along the Wabash River were significantly different (ANOVA: F 3,95 = 5.39, P = 0.002; Figure 2.5). Fish captured near rkm 499 were significantly smaller than those collected further upstream (rkm 521 and rkm 565), although the most upstream site (rkm 600) had a mean L T that was not significantly different from the other sampling locations. However, W R was not different among rkm sampled within 2011 (ANOVA: F 3,95 = 1.27, P > 0.05) with a mean W R of 1.09 ±0.01SE (n = 96).

30 Sex ratio Mean sex ratio for Silver Carp in the Wabash River was equal to 1 male for every 1.4 ±0.6SE females but did change among sampling months (Table 2.3). Females were more abundant in May 2011 as well as May and April 2013, but males were more abundant in May and September Males were also more abundant in June 2013; catch was evenly divided between sexes in July of the same year Length-at-maturation Length at 50% maturity for Silver Carp was 440 mm. The logistic regression model was also significant (Logistic regression: P 0.001), with accuracy of 96.6% (57 of 59 individuals successfully classified; Figure 2.6). The smallest mature female observed was 410 mm L T and the smallest mature male was 413 mm L T. 2.5 Discussion The impacts of invasive species can be far-reaching and variation in population characteristics among different ecosystems can help to evaluate the ecological impacts and future spread of these species. Wabash River Silver Carp varied in size structure, exhibiting decreasing mean L T and higher abundances of smaller individuals through time. Annual increases in smaller individuals have been observed in Illinois River Silver Carp (Phelps and Willis 2013), and decreases in mean length have also been documented in other invasive fishes (Bøhn et al. 2004; Amundsen et al. 2012; Feiner et al. 2012). Decreasing mean length can be influenced by reproduction via recruitment. Additionally, environmental differences (e.g., climate, competition) may influence growth and result in increasing abundances of smaller individuals. Invasive fishes are known to exhibit trends of declining mean length through time related to invasion stage and time since invasion

31 14 (Feiner et al. 2012). Temporal changes in population characteristics may also be analogous to increasing invasions stage and could indicate that Silver Carp populations in the Wabash River are still increasing. Condition (e.g., W R ) of Silver Carp in the Wabash River declined through time but did not change spatially. Annual changes in condition indicate the cause likely operates on an annual or seasonal temporal scale. Similar factors can influence length and condition in fishes and so condition may also be influenced by climate and competition. Although no long-term information on changes in Silver Carp densities in the Wabash River is available, a growing Silver Carp population would increase intraspecific competition and could cause the observed declines in W R. Additionally, changes in resource availability related to temporal changes in conditions could cause annual changes in condition (e.g., oxygen, temperature; Rätz and Lloret 2003; Casini et al. 2006). Drought can also influence fish condition (Matthews and Marsh-Matthews 2003), and the significant decline in condition observed in 2013 may have been influenced by the severe drought that occurred in Length-weight relationships in Wabash River Silver Carp were different between males and females and also changed annually. Females exhibited similar length-weight relationships in 2011 and 2012, but the length-weight relationship in 2013 had a steeper slope and smaller intercept. Small sample size in all but one year did not allow pre- and post-spawning females to be separated. However, female Silver Carp captured in all months contained eggs in some stage of development or reabsorption. The inclusion of females with eggs in calculating length-weight relationships likely influenced the steeper slopes observed in females compared to males in all but one year. Small sample size of

32 15 males may have contributed to the inverted relationships of males versus females in Slopes for females increased slightly through time but the slopes of male length-weight relationships were more variable length-weight regressions are likely the best representatives of this relationship in male and female Silver Carp from the Wabash River due to larger sample size. However, annual differences in length-weight relationship may be important indicators of annual changes in growth and condition within ecosystems. Spatial variation in Silver Carp L T within a year showed smaller individuals occurred downstream with the exception of the most upstream sampling location. The center two sampling locations had significantly greater mean L T than the other locations, although this finding may have been influenced by the small sample size at the most upstream site. Habitat selection, behavior, and dispersal can influence the sizes of individuals observed at different locations. In general, larger fish tend to move farther and can more easily disperse upstream than small fish (e.g.,albanese et al. 2004; Woolnough et al. 2008). Specific behaviors, such as territoriality, may alter dispersal across invaded ecosystems. For example, smaller Round Goby (Neogobius melanostomus) in the Laurentian Great Lakes were the first to colonize new areas (Coulter et al. 2012, Ray and Corkum 2001), but dispersers in an invaded river tended to be larger individuals (Brander et al. 2013). Overall, these findings suggest that smaller individuals may be on the leading edge of Silver Carp invasions, but given the small number of fishes collected at the most upstream location (3), additional sampling is required to solidify this statement.

33 16 The skewed sex ratios, both towards males and females, as well as length-atmaturation may have implications for the reproductive potential of Silver Carp in the Wabash River. Sex ratios skewed towards females can result in higher reproductive rates (Adel 2012; Russell et al. 2012). If the overall population of Silver Carp is skewed towards females, reproduction could increase in a given population. However, the observed Silver Carp sex ratio changed temporally and was sometimes skewed towards males. This variation indicates that the observed sex ratios in Wabash River may be caused by differences in movements (Hutchings and Gerber 2002; Croft et al. 2003) or habitat use (Darden and Croft 2008) which may vary seasonally. However, there was no clear trend, with May specifically alternating from female to male dominated from year to year. Plasticity in sex ratio may determine invasion success (Brandner et al. 2013), leading to increased reproduction during critical periods during establishment. Therefore, it will be important to investigate the cause and nature of the skewed sex ratios of Silver Carp in the largely free-flowing Wabash River to better understand how this may facilitate invasion by this species. Reproductive effort of a fish population can be further increased by early maturation exhibited by some populations of invasive fishes (e.g., MacInnis and Corkum 2000a; Fox et al. 2007; Amundsen et al. 2012), especially along invasion fronts. Silver Carp were found to have a 50% chance of being mature at a relatively small L T of 440 mm. In contrast, invasive Silver Carp populations from Russia and Japan achieved maturity at 600 mm or more (Costa-Pierce 1992) and at three to six years of age (Kolar et al. 2007). Based on Silver Carp length-at-age data from the lower Wabash River (Stuck et al. 2015), mature individuals in this study with L T of 440 mm may be as young as one or

34 17 two-years-old. Smaller L T could result from differences among ecosystems in climate, resource availability, and differences in both intra- and interspecific competition. However, length-at-maturity is also known to vary during the establishment and spread of invasive species. In recently introduced populations, smaller, younger individuals may mature faster compared to longer-established populations (Russell et al. 2012). Early maturation is linked with negative outcomes (e.g., Justus and Fox 1994), and so size and age at maturation may change plastically. As an invasive species, it is possible that Silver Carp may also exhibit plasticity in age or size at maturation, although additional information would be necessary to evaluate this possibility. Silver Carp present an interesting example of how the population characteristics of an invasive species may vary and exhibit interesting temporal variation in size structure, condition, and sex ratio. Additional information is still needed to examine all of the causes of this variation and to evaluate all possible traits that contribute to the success of Silver Carp as an invasive species. Traits that increase reproductive effort (e.g., fecundity, length-at-maturity, age-at-maturity) and could increase the likelihood of successful establishment (Bøhn et al. 2004; Copp et al. 2004; Copp and Fox 2007) may be especially important. Broad-scale changes in population and life history characteristics through invasion stages can also provide further insight into the reasons that a particular species may successfully establish and improve predictions and management of invasive species.

35 Figure 2.1. Temporal variation in electrofished Silver Carp total length (L T ±SE) from river kilometer 499 in the Wabash River, Indiana, USA. Letters indicate significant differences 18

36 Figure 2.2. Distributions of total lengths of electrofished Silver Carp from river kilometer 499 collected from in the Wabash River, Indiana, USA. 19

37 Figure 2.2. Temporal variation is electrofished Silver Carp relative weight (W R ±SE) from river kilometer 499 in the Wabash River, Indiana, USA. Letters indicate significant differences. 20